Linear Regression as the Simplest Classifier

Posted on Mon 24 February 2020 in Posts • Tagged with machine learning, linear regression, python, scikit-learn, statistical learning

In this post I wanted to describe a simple application of a linear least squares method to a problem of data classification. It is a naive approach and is unlikely to beat more sophisticated techniques like Logistic Regression, for instance.


Some imports we are going to need for this …

Continue reading

How to write a decent training loop with enough flexibility.

Posted on Sat 15 June 2019 in Posts • Tagged with deep learning

In this post, I briefly describe my experience in setting up training with PyTorch.


PyTorch is an extremely useful and convenient framework for deep learning. When it comes to working on a deep learning project, I am more comfortable with PyTorch rather than TensorFlow.

In this quick post, I …

Continue reading

RiCNN and Rotation Robustness of ConvNets. A Paper Review

Posted on Sat 15 June 2019 in Posts • Tagged with review series, deep learning, computer vision

Lately, I have been reading more papers on modern advances in deep learning in order to get a clear view of what problem I want to focus on during my PhD research.

There is a lot of information to process and an incredible amount of papers are being published from …

Continue reading

Computer Vision: Can You Teach a Machine To See?

Posted on Fri 22 March 2019 in Posts • Tagged with education, talks

A little overview of what I talked about at CUNY CSI Science Day.

During the Science Day at the CUNY College of Staten Island, I presented a gentle introduction to area of computer vision with fun examples and research results to visiting middle and high school students.

It was a …

Continue reading

Harmonic networks: implementation of paper results

Posted on Sun 10 March 2019 in Posts • Tagged with deep learning, computer vision, work in progress

I implement an interesting result from a recent paper on convolutional neural networks.


In this post I will briefly discuss my implementation of a model introduced in this paper.

In short, the authors suggest using predefined filters in a convolutional network based on Discrete Cosine Transform.

I used PyTorch …

Continue reading

Image Quality Measure

Posted on Thu 13 December 2018 in Posts • Tagged with image enhancement, image quality

A simple function that can be used to justify image quality and control enhancement.


One difficult thing about image enhancement is to actually measure the level of image quality which is quite a subjective task. On the one hand, each individual can perceive the image quality according to their …

Continue reading

A surprising way sigmoid function is applied in computer vision

Posted on Fri 30 November 2018 in Posts • Tagged with image enhancement

Let's talk about all things image enhancement, what it is, why it is necessary and how do wavelets play a big part in it!


Enhancement of images is an important preprocessing step in any image related system. Getting rid of noise, brightening, extraction of details - all of this helps …

Continue reading